Problem: Estimating $\ln(1.7)$ using a Taylor polynomial about $x=2$, what is the least degree of the polynomial that assures an error smaller than $0.001$ ? (The $n^{\text{th}}$ derivative of $\ln(x)$ is $(-1)^{n-1}\dfrac{(n-1)!}{x^n}$ for $n\geq1$ ) Choose 1 answer: Choose 1 answer: (Choice A) A $2$ (Choice B) B $3$ (Choice C) C $4$ (Choice D) D $5$
Solution: We will use the Lagrange error bound. Let's assume the polynomial's degree is $n$. The $(n+1)^{\text{th}}$ derivative of $\ln(x)$ is $(-1)^n\dfrac{n!}{x^{n+1}}$. The Lagrange bound for the error assures that $\begin{aligned} |R_n(1.7)|&\leq\left|\dfrac{(-1)^n\dfrac{n!}{z^{n+1}}}{(n+1)!}(1.7-2)^{n+1}\right| \\\\ &=\dfrac{n!\cdot 0.3^{n+1}}{(n+1)!z^{n+1}} \\\\ &=\dfrac{0.3^{n+1}}{(n+1)z^{n+1}} \end{aligned}$ where $1.7\leq z\leq 2$. On this interval, $\dfrac{0.3^{n+1}}{(n+1)z^{n+1}}\leq\dfrac{0.3^{n+1}}{(n+1)1.7^{n+1}}$. Solving $\dfrac{0.3^{n+1}}{(n+1)1.7^{n+1}}<0.001$ using trial and error, we find that $n\geq3$. In conclusion, the least degree of the polynomial that assures our error bound is $3$.